CIRP Annals - Manufacturing Technology

Transcription

1 CIRP Annals - Manufacturing Technology 59 (2010) Contents lists available at ScienceDirect CIRP Annals - Manufacturing Technology journal homepage: Investigations on the effects of multi-layered coated inserts in machining Ti 6Al 4V alloy with experiments and finite element simulations T. Özel (2) a, *, M. Sima a, A.K. Srivastava (3) b, B. Kaftanoglu (1) c a Manufacturing Automation Research Laboratory, Department of Industrial and Systems Engineering, Rutgers University, NJ, USA b TechSolve Inc., Cincinnati, OH, USA c Department of Manufacturing Engineering, Atilim University, Ankara, Turkey ARTICLE INFO ABSTRACT Keywords: Machining Finite element method Titanium Tool This paper presents investigations on turning Ti 6Al 4V alloy with multi-layer coated inserts. Turning of Ti 6Al 4V using uncoated, TiAlN coated, and TiAlN + cbn coated single and multi-layer coated tungsten carbide inserts is conducted, forces and tool wear are measured. 3D finite element modelling is utilized to predict chip formation, forces, temperatures and tool wear on these inserts. Modified material models with strain softening effect are developed to simulate chip formation with finite element analysis and investigate temperature fields for coated inserts. Predicted forces and tool wear contours are compared with experiments. The temperature distributions and tool wear contours demonstrate some advantages of coated insert designs. ß 2010 CIRP. 1. Introduction Titanium alloys such as Ti 6Al 4V offer high strength-toweight ratio, high toughness, superb corrosion resistance, and biocompatibility and are increasingly used in aerospace and biomedical applications. However, titanium alloys are difficult to machine due to their low thermal conductivity and diffusivity, high rigidity and low elasticity modulus and high chemical reactivity at elevated temperatures [1]. These alloys exhibit serrated and cyclical chip formation resulting in detrimental tool vibrations [2]. Titanium alloy machining performance can be increased by improving cutting tool materials and coatings [3,4]. Cubic boron nitride (cbn) material offers outstanding properties such as high hardness and wear resistance. However, cbn material has lower toughness and is not suitable for forming inserts into complex shapes. Recently, cbn coatings have been explored by applying several deposition techniques. Among these, physical vapor deposition (PVD) has been preferred since thinner coatings can be deposited and sharp edges and complex shapes can be easily coated at lower temperatures. On the other hand, coating applied affects the edge radius of the inserts and must be taken into consideration during tool performance analysis [4]. In this research, single and multi-layer TiAlN and cbn coatings are experimented on tungsten carbide inserts (WC/Co) for possible improvements in machining of Ti 6Al 4V alloy. Positive tool geometry WC/Co inserts are coated by cbn using a PVD system. Machining performance of multi-layered coated inserts is examined in longitudinal turning of titanium alloy (Ti 6Al 4V) without using coolant. The performance of uncoated, TiAlN coated, cbn * Corresponding author. coated and multi-layer cbn + TiAlN coated tungsten carbide inserts is compared in terms of cutting forces and tool wear. In addition, finite element (FE) simulations are utilized in investigating the tool temperatures and wear development. Two dimensional and three dimensional FE simulations have been designed and conducted to predict forces, temperatures and tool wear to investigate the advantages of coatings in machining of Ti 6Al 4V. Finite element modelling and simulations of chip formation process in titanium alloy machining present significant challenges due to the nature of complicated dynamic material behaviour of these alloys at elevated temperatures, strain and strain rates. In general, adiabatic shearing is considered as responsible for serrated chip formation. Increasing temperatures in the primary shear zone due to shear deformation weaken the material by thermal softening; therefore, the deformation is concentrated in shear bands, leading to serrated chip formation [5]. Although it is also possible to simulate serrated chip formation by using damage models [6], in this paper it is assumed that serration is caused by adiabatic shearing. 2. Material constitutive model for Ti 6Al 4V and validation In FE models, a constitutive material model is required to relate the flow stress to strain, strain rate and temperature, which often obtained from Split-Hopkinson pressure bar (SHPB) tests performed under various strain rates and temperatures. Dynamic material behaviour for Ti 6Al 4V titanium alloy has been widely published in literature [7,8]. Nemat-Nasser [7] reported that a phenomenon known as strain (flow) softening is observed which is responsible for adiabatic shearing in titanium alloys. Localized softening is described as offering less resistance to local deformations due to rearrangement of dislocations caused by subsequent cycling in hard materials. This phenomenon is usually seen during /$ see front matter ß 2010 CIRP. doi: /j.cirp

2 78 T. Özel et al. / CIRP Annals - Manufacturing Technology 59 (2010) an increase in strain beyond a critical strain value. Specifically, Lee and Lin [8] investigated temperature and strain-rate sensitivity of Ti 6Al 4V and presented flow stress data at temperatures from 20 to C and strain rates ranging from 800 up to 3300 s 1. They used The Johnson-Cook (JC) material model to represent their flow stress data. However, their model did not include temperaturedependent strain softening effect Modified material model A modification to the JC model is offered to include strain (flow) softening effects at elevated temperatures as proposed by Calamaz et al. [9]. In this study, further modifications to the strain hardening part of the JC model by including strain softening at higher strain values and thermal softening part are proposed and the model is given in Eq. (1). In this model, the influence of strain, strain rate, temperature and temperature-dependent strain softening on the flow stress is defined by four multiplicative terms. s ¼ A þ Be n 1 expðe a 1 þ C ln ė Þ ė 0 1 T T m 0 1 s (1) D þð1 DÞ tan h T m T 0 ðe þ pþ r where D ¼ 1 ðt=t m Þ d ; p ¼ðT=T m Þ b, s is flow stress, e is true strain, ė is true strain rate, ė 0 is reference true strain, and T, T m,t 0 are work, material melting and ambient temperatures respectively. The experimental flow stress data by Lee and Lin [8] has been taken as the base for this modified material model. The most optimum set of model parameters that was identified with inverse analysis are; A = MPa, B = MPa, n = 0.28, C = 0.028, m = 1.0, a =2, s = 0.05, r =2, d =1, b = 5. The details of this methodology are outlined in the work by Özel et al. [10] Orthogonal cutting tests Orthogonal turning of Ti 6Al 4V titanium alloy tubes (50.8 mm in diameter and mm in thickness) have been performed using uncoated and TiAlN coated tungsten carbide (WC/Co) cutting tools in a rigid CNC turning centre at TechSolve Inc. The cutting forces were measured with a force dynamometer and high-speed data acquisition devices. The experiments have been conducted using tool holders with 08 and 58 rake angle (g) at a cutting speed of v c = 120 m/min and three different feeds (f = 0.075, 0.1, mm/ rev). Images of micro-chip geometries were captured with optical digital microscopy D finite element simulations and validation Finite element model is developed using updated Lagrangian (DEFORM-2D) software in which chip separation from workpiece is achieved with continuous remeshing. A plane-strain coupled thermo-mechanical analysis was performed. In these simulations, serrated chip formation process is simulated from the incipient to the steady-state by using adiabatic shearing based on strain (flow) softening elasto-viscoplastic work material assumption. The Table 1 Mechanical and thermo-physical properties of work and tool materials and friction values used in FE simulations. Ti 6Al 4V WC/Co (Ti,Al)N cbn E(T) [MPa] *T e 5 6.0e e 5 a(t) [mm mm 1 8C 1 ] *T e 6 9.4e 6 5.2e 6 l(t) [Wm 1 8C 1 ] 7.039e *T *T c p (T) [Nmm 2 8C 1 ] 2.24e *T *T *T Fig. 1. Comparison of measured and simulated serrated chip geometry.

3 T. Özel et al. / CIRP Annals - Manufacturing Technology 59 (2010) Table 2 Comparison of measured and simulated chip geometry. Cutting condition Experimental Simulated Tool g [8] f [mm/rev] h 1 [mm] h 2 [mm] h 1 [mm] h 2 [mm] WC/Co WC/Co WC/Co WC/Co WC/Co + TiAlN WC/Co + TiAlN WC/Co + TiAlN WC/Co + TiAlN simulations included a workpiece as elasto-viscoplastic with a mesh containing 10,000 quadrilateral elements. Tool is modelled as rigid with a mesh containing 2500 elements. Rake angles of 08 and 58 are employed in the tool geometry. Tool edge radius was estimated to be r b =5mm for uncoated WC/Co, and r b =10mm for TiAlN coated WC/Co respectively. Thermal boundary conditions are defined accordingly in order to allow heat transfer from workpiece to the cutting tool. The heat conduction coefficient (h) is taken as 1.0e 5 kw m 2 K 1 to allow rapid temperature rise in the tool. Mechanical and thermo-physical properties of titanium Ti 6Al 4V alloy are defined as temperature (T) dependent. Temperature-dependent (T in 8C) modulus of elasticity (E), thermal expansion (a), thermal conductivity (l), and heat capacity (c p ) are given in Table 1. In this paper, three contact regions are considered at the tool chip interface: (i) a sticking region from the tool tip point to the end of the round edge curvature (t = k where t is frictional shear stress and k is the work material shear flow stress), (ii) a shear friction region (m = t/k) from the end of the curvature to the uncut chip thickness boundary (m = 0.9 for WC/Co, m = 0.85 for TiAlN), (iii) a sliding region along the rest of the rake face (m = 0.7 for WC/Co, m = 0.5 for TiAlN as the friction coefficient). Simulations are run for 0.1 s cutting time. Comparison of simulated chips with experiments that are shown in Fig. 1 and Table 2 indicate close agreements. In Table 2, h 1 and h 2 indicate minimum and maximum serrated chip thickness respectively. Predicted forces from simulations are compared with measured forces in orthogonal cutting tests of Ti 6Al 4V alloy tubes as Fig. 2. Comparison of measured and simulated forces in orthogonal cutting tests.

4 80 T. Özel et al. / CIRP Annals - Manufacturing Technology 59 (2010) Fig. 3. Configuration of longitudinal bar turning experiments and force results. shown in Fig. 2. Especially, cutting forces are in close agreements with 5% prediction error. Thrust force predictions which show 10 15% prediction error can be further improved with finer adjustments of friction regions and their values. 3. Experimental work In this study, four different inserts at the same cutting conditions have been tested; uncoated/unalloyed tungsten carbide (WC/Co), tungsten carbide (WC/Co) PVD coated with TiAlN, tungsten carbide (WC/Co) PVD coated with cbn, tungsten carbide multi-layer PVD coated with cbn over TiAlN coating. Tungsten carbide (WC/Co) and PVD coated WC/Co with TiAlN inserts are coated with cbn by magnetron sputtering PVD system as mono and multi-layer coatings at National Boron Research Institute (BOREN) in Turkey at a deposition pressure of Torr and heater temperature of 100 8C. Applied magnetron power is fixed at 900 W and argon to nitrogen gas ratio is adjusted to 5/1 and run at the lowest possible bias voltage to obtain uniform cbn coating. Longitudinal turning of annealed Ti 6Al 4V titanium alloy bars (90 mm in diameter, 100 mm in length) was performed by using TPG432 type insert geometry (insert nose radius of r e = 0.8 mm and relief angle of a =118) in a rigid CNC turning centre under dry machining conditions at Rutgers University Manufacturing Automation Research Laboratory. The inserts were used with a tool holder that provided 08 lead, 58 side rake, and 58 back rake angles. The cutting forces were measured with a force dynamometer mounted on the turret disk of the CNC turning centre. A Table 3 Summary of FE simulation predictions. Tool type F c [N] F t [N] F z [N] T tool [8C] T chip [8C] dw/dt [mm/s] WC/C ( ) 93 (69 106) 229 ( ) WC/C 0 + cbn 602 ( ) 99 (86 100) 236 ( ) WC/C 0 + TiAlN 571 ( ) 98 (77 103) 228 ( ) WC/C 0 + TiAlN + cbn 575 ( ) 97 (81 115) 243 ( ) Fig. 4. Predicted temperature distributions in 8C.

5 T. O zel et al. / CIRP Annals - Manufacturing Technology 59 (2010) Fig. 5. Simulated chip formation with effective strain distributions. constant depth of cut (ap = 2 mm) and a constant feed (f = 0.1 mm/ rev) were selected as cutting conditions. Each test was replicated at least twice. The averages of the measured forces for each insert are shown in Fig. 3. In order to observe the performance of coatings at different cutting speeds, two sets of tests are done at cutting speeds of vc = 50 and 100 m/min respectively. Thrust force was the lowest since inserts use 118 relief angle; hence flank contact area is very Fig. 6. Experimental and predicted wear rate distributions in [mm/s].

6 82 T. Özel et al. / CIRP Annals - Manufacturing Technology 59 (2010) small. According to force measurements, cbn and cbn + TiAlN coated inserts exhibit lowest cutting forces at 50 m/min cutting speed but the highest at 100 m/min cutting speed. Moreover, the highest thrust force is seen in cbn coated WC/Co inserts at high cutting speed. The advantage of cbn coatings on forces is apparent for the lower cutting speed. Adding cbn coating over TiAlN coating decreases forces. As cutting speed increases, the effect of larger edge radius (r b ) due to added layer of coatings (cbn and TiAlN) becomes the dominant mechanism on forces. This larger edge radius in multi-layer coated tools hinders the potential benefits of coatings, hence results in higher forces especially when cutting speed is doubled. Hence, it may be beneficial to modify edge preparation of the coated tools (TiAlN and cbn) to lower the cutting edge radius. 4. 3D finite element simulations Several FE studies on 3D turning are presented in the past such as the work by Aurich and Bil [11] for segmented chip formation. In this study, updated Lagrangian FE modelling software (DEFORM- 3D) was used. The workpiece is modelled as elastic viscoplastic material where the material constitutive model of this deformable body is represented with modified J-C material model. The workpiece is represented by a curved model with 87 mm diameter which is consistent with the experimental conditions. Only a segment (38) of the workpiece was modelled in order to keep the size of mesh elements small. Workpiece model includes 90,000 elements. The bottom surface of the workpiece is fixed in all directions. The cutting tool (r e = 0.8 mm with 118 relief angle) is modelled as a rigid body which moves at the specified cutting speed by using 180,000 elements. A very fine mesh density is defined at the tip of the tool and at the cutting zone to obtain fine process output distributions. The minimum element size for the workpiece and tool mesh was set to and mm respectively. A tool edge radius of 5, 10 and 15 mm are designed for uncoated, single layer and multi-layer coated tools respectively for each simulation, since added layers in multi-coating design is increasing the edge radius of the inserts [4]. All simulations were run at the same experimental cutting condition (v c = 100 m/min, f = 0.1 mm/rev, a p = 2 mm). In 3D FE modelling, constant shear friction factor (m = ) was used to represent friction between tool and workpiece. The averages, minimum and maximum of the simulated forces (F c, cutting force, F t, thrust force and F z, feed force) are given in Table 3. The simulated cutting forces are found to be in close agreements with the experimental ones as summarized in Table 3. Predicted temperature distributions (see Fig. 4) depict that the lowest temperature rise in the tool is observed with cbn coated WC/Co tool due to the highest effective thermal conductivity and contact friction. Predicted chip formation is shown in Fig. 5. 3D FE simulations are also utilized to predict tool wear. The tool wear rate models describe the rate of volume loss on the tool rake and flank faces per unit area per unit time. A tool wear rate model based on the adhesive wear proposed by Usui et al. [12] was employed. This model uses interface temperature (T), normal stress (s n ), and sliding velocity (v s ) at the contact surfaces as inputs and yields tool wear rate (dw/dt) for a given location on the tool surface as shown in Eq. (2). The constants of this equation are set to c 1 = and c 2 = in the FE simulations. Determining these constants could be extensive work so they are kept same for simplicity. Chiefly crater wear was observed on all of the tools under these cutting conditions. Fig. 6 shows the comparison of measured and simulated tool wear zones. dw=dt ¼ c 1 s n y s exp c 2 (2) T 5. Conclusions In this study, a modified material model for Ti 6Al 4V titanium alloy is developed where strain (flow) softening, strain hardening and thermal softening effects are coupled. This model is validated with elasto-viscoplastic FE simulation of adiabatic shearing based serrated chip formation in machining Ti 6Al 4V titanium alloy. The simulation predictions are compared with orthogonal cutting test results by using measured forces and chip morphology. Turning experiments have been conducted with uncoated, mono and multi-layer coated WC/Co carbide tools and cutting performance of these coatings are evaluated. Although cbn and TiAlN + cbn coated WC/CO inserts exhibit largest cutting forces at higher cutting speeds, they reveal favourable wear development. Tool wear zone measurements and predictions show that cbn coated WC/Co inserts depict smallest wear zone. Consequently, cbn coatings may lead to reduction in tool wear dry machining of titanium alloyed Ti 6Al 4V material. Acknowledgements The authors gratefully acknowledge the support by the National Science Foundation (CMMI and CMMI ), and BOREN Institute of Turkey. References [1] Byrne G, Dornfeld D, Denkena B (2003) Advanced Cutting Technology. CIRP Annals 52(2): [2] Cotterell M, Byrne G (2008) Dynamics of Chip Formation during Orthogonal Cutting of Titanium Alloy Ti 6Al 4V. CIRP Annals 57: [3] Corduan N, et al, (2003) Wear Mechanisms of New Tool Materials for Ti 6AI- 4V High Performance Machining. CIRP Annals 52(1): [4] Bouzakis K-D, et al, (2009) Application in Milling of Coated Tools with Rounded Cutting Edges after Film Deposition. CIRP Annals 58(1): [5] Komanduri R, von Turkovich BF (1981) New Observations on the Mechanism of Chip Formation When Machining Titanium Alloys. Wear 69: [6] Shivpuri R, Hua J, Mittal P, Srivastava AK (2002) Microstructure-Mechanics Interactions in Modelling Chip Segmentation During Titanium Machining. CIRP Annals 51(1): [7] Nemat-Nasser S, et al, (2001) Dynamic Response of Conventional and Hot Isostatically Pressed Ti 6Al 4V Alloys: Experiments and Modelling. Mechanics of Materials 33: [8] Lee WS, Lin CF (1998) Plastic Deformation and Fracture Behaviour of Ti6Al4V Alloy Loaded with High Strain Rate Under Various Temperatures. Materials Science and Engineering A 241: [9] Calamaz M, Coupard D, Girot F (2008) A New Material Model for 2D Numerical Simulation of Serrated Chip Formation When Machining Titanium Alloy. Ti 6Al 4V. International Journal of Machine Tools and Manufacture 48: [10] Özel T, Sima M, Srivastava AK (2010) Finite Element Simulation of High Speed Machining Ti 6Al 4V Alloy using Modified Material Models. Transactions of the NAMRI/SME 38:1 8. [11] Aurich JC, Bil H (2006) 3D Finite Element Modelling of Segmented Chip Formation. CIRP Annals 55(1): [12] Usui E, Hirota A, Masuko M (1978) Analytical Prediction of Three Dimensional Cutting Process, Part 3: Cutting Temperature and Crater Wear of Carbide Tool. Journal of Engineering for Industry 100: